refactor(traverse): separate #[scope] attr (#3901)

Separate out attributes which communicate info to codegen related to scopes into `#[scope]` attr.

Before:

```rs
#[visited_node(scope(ScopeFlags::empty()))]
pub struct BlockStatement<'a> { /* ... */ }
```

After:

```rs
#[visited_node]
#[scope]
pub struct BlockStatement<'a> { /* ... */ }
```

I think this is clearer.
This commit is contained in:
overlookmotel 2024-06-26 05:43:09 +00:00
parent 5ef28b7375
commit 1061baabbf
5 changed files with 172 additions and 96 deletions

View file

@ -1,3 +1,6 @@
// NB: `#[visited_node]` and `#[scope]` attributes on AST nodes do not do anything to the code in this file.
// They are purely markers for codegen used in `oxc_traverse`. See docs in that crate.
// Silence erroneous warnings from Rust Analyser for `#[derive(Tsify)]`
#![allow(non_snake_case)]
@ -23,8 +26,9 @@ use serde::Serialize;
#[cfg(feature = "serialize")]
use tsify::Tsify;
#[visited_node(
scope(ScopeFlags::Top),
#[visited_node]
#[scope(
flags(ScopeFlags::Top),
strict_if(self.source_type.is_strict() || self.directives.iter().any(Directive::is_use_strict)),
)]
#[derive(Debug)]
@ -943,7 +947,8 @@ pub struct Hashbang<'a> {
}
/// Block Statement
#[visited_node(scope(ScopeFlags::empty()))]
#[visited_node]
#[scope]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1099,10 +1104,8 @@ pub struct WhileStatement<'a> {
}
/// For Statement
#[visited_node(
scope(ScopeFlags::empty()),
scope_if(self.init.as_ref().is_some_and(ForStatementInit::is_lexical_declaration)),
)]
#[visited_node]
#[scope(if(self.init.as_ref().is_some_and(ForStatementInit::is_lexical_declaration)))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1136,7 +1139,8 @@ pub enum ForStatementInit<'a> {
}
/// For-In Statement
#[visited_node(scope(ScopeFlags::empty()), scope_if(self.left.is_lexical_declaration()))]
#[visited_node]
#[scope(if(self.left.is_lexical_declaration()))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1168,7 +1172,8 @@ pub enum ForStatementLeft<'a> {
}
}
/// For-Of Statement
#[visited_node(scope(ScopeFlags::empty()), scope_if(self.left.is_lexical_declaration()))]
#[visited_node]
#[scope(if(self.left.is_lexical_declaration()))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1228,7 +1233,8 @@ pub struct WithStatement<'a> {
}
/// Switch Statement
#[visited_node(scope(ScopeFlags::empty()))]
#[visited_node]
#[scope]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1288,7 +1294,8 @@ pub struct TryStatement<'a> {
pub finalizer: Option<Box<'a, BlockStatement<'a>>>,
}
#[visited_node(scope(ScopeFlags::empty()), scope_if(self.param.is_some()))]
#[visited_node]
#[scope(if(self.param.is_some()))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -1422,9 +1429,10 @@ pub struct BindingRestElement<'a> {
}
/// Function Definitions
#[visited_node(
#[visited_node]
#[scope(
// TODO: `ScopeFlags::Function` is not correct if this is a `MethodDefinition`
scope(ScopeFlags::Function),
flags(ScopeFlags::Function),
strict_if(self.body.as_ref().is_some_and(|body| body.has_use_strict_directive())),
)]
#[derive(Debug)]
@ -1530,8 +1538,9 @@ pub struct FunctionBody<'a> {
}
/// Arrow Function Definitions
#[visited_node(
scope(ScopeFlags::Function | ScopeFlags::Arrow),
#[visited_node]
#[scope(
flags(ScopeFlags::Function | ScopeFlags::Arrow),
strict_if(self.body.has_use_strict_directive()),
)]
#[derive(Debug)]
@ -1565,7 +1574,8 @@ pub struct YieldExpression<'a> {
}
/// Class Definitions
#[visited_node(scope(ScopeFlags::StrictMode))]
#[visited_node]
#[scope(flags(ScopeFlags::StrictMode))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(rename_all = "camelCase"))]
@ -1690,7 +1700,8 @@ pub struct PrivateIdentifier<'a> {
pub name: Atom<'a>,
}
#[visited_node(scope(ScopeFlags::ClassStaticBlock))]
#[visited_node]
#[scope(flags(ScopeFlags::ClassStaticBlock))]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]

View file

@ -1,7 +1,7 @@
//! [JSX](https://facebook.github.io/jsx)
// NB: `#[visited_node]` attribute on AST nodes does not do anything to the code in this file.
// It is purely a marker for codegen used in `oxc_traverse`. See docs in that crate.
// NB: `#[visited_node]` and `#[scope]` attributes on AST nodes do not do anything to the code in this file.
// They are purely markers for codegen used in `oxc_traverse`. See docs in that crate.
// Silence erroneous warnings from Rust Analyser for `#[derive(Tsify)]`
#![allow(non_snake_case)]

View file

@ -1,7 +1,7 @@
//! Literals
// NB: `#[visited_node]` attribute on AST nodes does not do anything to the code in this file.
// It is purely a marker for codegen used in `oxc_traverse`. See docs in that crate.
// NB: `#[visited_node]` and `#[scope]` attributes on AST nodes do not do anything to the code in this file.
// They are purely markers for codegen used in `oxc_traverse`. See docs in that crate.
// Silence erroneous warnings from Rust Analyser for `#[derive(Tsify)]`
#![allow(non_snake_case)]

View file

@ -3,8 +3,8 @@
//! [AST Spec](https://github.com/typescript-eslint/typescript-eslint/tree/main/packages/ast-spec)
//! [Archived TypeScript spec](https://github.com/microsoft/TypeScript/blob/3c99d50da5a579d9fa92d02664b1b66d4ff55944/doc/spec-ARCHIVED.md)
// NB: `#[visited_node]` attribute on AST nodes does not do anything to the code in this file.
// It is purely a marker for codegen used in `oxc_traverse`. See docs in that crate.
// NB: `#[visited_node]` and `#[scope]` attributes on AST nodes do not do anything to the code in this file.
// They are purely markers for codegen used in `oxc_traverse`. See docs in that crate.
// Silence erroneous warnings from Rust Analyser for `#[derive(Tsify)]`
#![allow(non_snake_case)]
@ -46,7 +46,8 @@ pub struct TSThisParameter<'a> {
/// Enum Declaration
///
/// `const_opt`enum`BindingIdentifier`{`EnumBody_opt`}
#[visited_node(scope(ScopeFlags::empty()))]
#[visited_node]
#[scope]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type"))]
@ -562,7 +563,8 @@ pub struct TSTypeParameterInstantiation<'a> {
pub params: Vec<'a, TSType<'a>>,
}
#[visited_node(scope(ScopeFlags::empty()))]
#[visited_node]
#[scope]
#[derive(Debug)]
#[cfg_attr(feature = "serialize", derive(Serialize, Tsify))]
#[cfg_attr(feature = "serialize", serde(tag = "type", rename_all = "camelCase"))]
@ -783,8 +785,9 @@ pub enum TSTypePredicateName<'a> {
This(TSThisType),
}
#[visited_node(
scope(ScopeFlags::TsModuleBlock),
#[visited_node]
#[scope(
flags(ScopeFlags::TsModuleBlock),
strict_if(self.body.as_ref().is_some_and(|body| body.is_strict())),
)]
#[derive(Debug)]

View file

@ -1,8 +1,7 @@
import {readFile} from 'fs/promises';
import {join as pathJoin} from 'path';
import {fileURLToPath} from 'url';
import assert from 'assert';
import {typeAndWrappers, snakeToCamel} from './utils.mjs';
import {typeAndWrappers} from './utils.mjs';
const FILENAMES = ['js.rs', 'jsx.rs', 'literal.rs', 'ts.rs'];
@ -20,70 +19,113 @@ export default async function getTypesFromCode() {
return types;
}
class Position {
constructor(filename, index) {
this.filename = filename;
this.index = index;
}
assert(condition, message) {
if (!condition) this.throw(message);
}
throw(message) {
throw new Error(`${message || 'Unknown error'} (at ${this.filename}:${this.index + 1})`);
}
}
class Lines {
constructor(lines, filename, offset = 0) {
this.lines = lines;
this.filename = filename;
this.offset = offset;
this.index = 0;
}
static fromCode(code, filename) {
const lines = code.split(/\r?\n/)
.map(line => line.replace(/\s+/g, ' ').replace(/ ?\/\/.*$/, '').replace(/ $/, ''));
return new Lines(lines, filename, 0);
}
child() {
return new Lines([], this.filename, this.index);
}
current() {
return this.lines[this.index];
}
next() {
return this.lines[this.index++];
}
isEnd() {
return this.index === this.lines.length;
}
position() {
return new Position(this.filename, this.index + this.offset);
}
positionPrevious() {
return new Position(this.filename, this.index + this.offset - 1);
}
}
function parseFile(code, filename, types) {
const lines = code.split(/\r?\n/).map(
line => line.replace(/\s+/g, ' ').replace(/ ?\/\/.*$/, '')
);
for (let lineIndex = 0; lineIndex < lines.length; lineIndex++) {
const lineMatch = lines[lineIndex].match(/^#\[visited_node ?([\]\(])/);
if (!lineMatch) continue;
let scopeArgs = null;
if (lineMatch[1] === '(') {
let line = lines[lineIndex].slice(lineMatch[0].length),
scopeArgsStr = '';
while (!line.endsWith(')]')) {
scopeArgsStr += ` ${line}`;
line = lines[++lineIndex];
}
scopeArgsStr += ` ${line.slice(0, -2)}`;
scopeArgsStr = scopeArgsStr.trim().replace(/ +/g, ' ').replace(/,$/, '');
scopeArgs = parseScopeArgs(scopeArgsStr, filename, lineIndex);
const lines = Lines.fromCode(code, filename);
while (!lines.isEnd()) {
if (lines.current() !== '#[visited_node]') {
lines.next();
continue;
}
let match;
while (true) {
match = lines[++lineIndex].match(/^pub (enum|struct) ((.+?)(?:<'a>)?) \{/);
// Consume attrs and comments, parse `#[scope]` attr
let match, scopeArgs = null;
while (!lines.isEnd()) {
if (/^#\[scope[(\]]/.test(lines.current())) {
scopeArgs = parseScopeArgs(lines, scopeArgs);
continue;
}
match = lines.next().match(/^pub (enum|struct) ((.+?)(?:<'a>)?) \{/);
if (match) break;
}
const [, kind, rawName, name] = match,
startLineIndex = lineIndex;
lines.position().assert(match, `Could not find enum or struct after #[visited_node]`);
const [, kind, rawName, name] = match;
const itemLines = [];
while (true) {
const line = lines[++lineIndex].trim();
// Find end of struct / enum
const itemLines = lines.child();
while (!lines.isEnd()) {
const line = lines.next();
if (line === '}') break;
if (line !== '') itemLines.push(line);
itemLines.lines.push(line.trim());
}
if (kind === 'struct') {
types[name] = parseStruct(name, rawName, itemLines, scopeArgs, filename, startLineIndex);
types[name] = parseStruct(name, rawName, itemLines, scopeArgs);
} else {
types[name] = parseEnum(name, rawName, itemLines, filename, startLineIndex);
types[name] = parseEnum(name, rawName, itemLines);
}
}
}
function parseStruct(name, rawName, lines, scopeArgs, filename, startLineIndex) {
function parseStruct(name, rawName, lines, scopeArgs) {
const fields = [];
for (let i = 0; i < lines.length; i++) {
let line = lines[i];
const isScopeEntry = line === '#[scope(enter_before)]';
if (isScopeEntry) {
line = lines[++i];
} else if (line.startsWith('#[')) {
while (!lines[i].endsWith(']')) {
i++;
while (!lines.isEnd()) {
let isScopeEntry = false, line;
while (!lines.isEnd()) {
line = lines.next();
if (line === '') continue;
if (line === '#[scope(enter_before)]') {
isScopeEntry = true;
} else if (line.startsWith('#[')) {
while (!line.endsWith(']')) {
line = lines.next();
}
} else {
break;
}
continue;
}
const match = line.match(/^pub ((?:r#)?([a-z_]+)): (.+),$/);
assert(
match,
`Cannot parse line ${startLineIndex + i} in '${filename}' as struct field: '${line}'`
);
lines.positionPrevious().assert(match, `Cannot parse line as struct field: '${line}'`);
const [, rawName, name, rawTypeName] = match,
typeName = rawTypeName.replace(/<'a>/g, '').replace(/<'a, ?/g, '<'),
{name: innerTypeName, wrappers} = typeAndWrappers(typeName);
@ -95,10 +137,19 @@ function parseStruct(name, rawName, lines, scopeArgs, filename, startLineIndex)
return {kind: 'struct', name, rawName, fields, scopeArgs};
}
function parseEnum(name, rawName, lines, filename, startLineIndex) {
function parseEnum(name, rawName, lines) {
const variants = [],
inherits = [];
for (const [lineIndex, line] of lines.entries()) {
while (!lines.isEnd()) {
let line = lines.next();
if (line === '') continue;
if (line.startsWith('#[')) {
while (!line.endsWith(']')) {
line = lines.next();
}
continue;
}
const match = line.match(/^(.+?)\((.+?)\)(?: ?= ?(\d+))?,$/);
if (match) {
const [, name, rawTypeName, discriminantStr] = match,
@ -108,34 +159,50 @@ function parseEnum(name, rawName, lines, filename, startLineIndex) {
variants.push({name, typeName, rawTypeName, innerTypeName, wrappers, discriminant});
} else {
const match2 = line.match(/^@inherit ([A-Za-z]+)$/);
assert(
match2,
`Cannot parse line ${startLineIndex + lineIndex} in '${filename}' as enum variant: '${line}'`
);
lines.positionPrevious().assert(match2, `Cannot parse line as enum variant: '${line}'`);
inherits.push(match2[1]);
}
}
return {kind: 'enum', name, rawName, variants, inherits};
}
function parseScopeArgs(argsStr, filename, lineIndex) {
if (!argsStr) return null;
function parseScopeArgs(lines, scopeArgs) {
const position = lines.position();
// Get whole of `#[scope]` attr text as a single line string
let scopeArgsStr = '';
let line = lines.next();
if (line !== '#[scope]') {
line = line.slice('#[scope('.length);
while (!line.endsWith(')]')) {
scopeArgsStr += ` ${line}`;
line = lines.next();
}
scopeArgsStr += ` ${line.slice(0, -2)}`;
scopeArgsStr = scopeArgsStr.trim().replace(/ +/g, ' ').replace(/,$/, '');
}
// Parse attr
return parseScopeArgsStr(scopeArgsStr, scopeArgs, position);
}
function parseScopeArgsStr(argsStr, args, position) {
if (!args) args = {flags: 'ScopeFlags::empty()', if: null, strictIf: null};
if (!argsStr) return args;
const matchAndConsume = (regex) => {
const match = argsStr.match(regex);
assert(match);
position.assert(match);
argsStr = argsStr.slice(match[0].length);
return match.slice(1);
};
const args = {};
try {
while (true) {
const [key] = matchAndConsume(/^([a-z_]+)\(/);
assert(
['scope', 'scope_if', 'strict_if'].includes(key),
`Unexpected visited_node macro arg: ${key}`
);
let [key] = matchAndConsume(/^([a-z_]+)\(/);
key = key.replace(/_([a-z])/g, (_, c) => c.toUpperCase());
position.assert(Object.hasOwn(args, key), `Unexpected scope macro arg: ${key}`);
let bracketCount = 1,
index = 0;
@ -148,21 +215,16 @@ function parseScopeArgs(argsStr, filename, lineIndex) {
if (bracketCount === 0) break;
}
}
assert(bracketCount === 0);
position.assert(bracketCount === 0);
const camelKey = key.replace(/_([a-z])/g, (_, c) => c.toUpperCase());
args[camelKey] = argsStr.slice(0, index).trim();
args[key] = argsStr.slice(0, index).trim();
argsStr = argsStr.slice(index + 1);
if (argsStr === '') break;
matchAndConsume(/^ ?, ?/);
}
assert(args.scope, 'Missing key `scope`');
} catch (err) {
throw new Error(
`Cannot parse visited_node args: ${argsStr} in ${filename}:${lineIndex}\n${err?.message}`
);
position.throw(`Cannot parse scope args: '${argsStr}': ${err?.message || 'Unknown error'}`);
}
return args;